​​[NOTES] Upcoming Traits Work

​​Possible path

​​improve higher-ranked types, lifetimes (needs issue)

​​Reason: 
  • ​​Chalk adopted an improved way of handling higher-ranked terms
  • ​​This will fix some bugs in the existing system (e.g., #32330) but also allows us to be much more expressive
  • ​​Needed, in particular, to support Generic Associated Types (GATs)
  • ​​Much work has been done here, but there is more to do
​​
​​Total estimate: 84 person-hours
​​
​​Steps:
  • ​​add a way to represent bound types in rustc
  • ​​Estimate: 40 person-hours
  • ​​prepare variant of lexical solver using universes instead
  • ​​Estimate: 24 person-hours
  • ​​integrate universes into canonicalization (48696)
  • ​​Estimate: 24 person-hours

​​integrate chalk under a flag (48049)

​​Reason:
  • ​​Gateway to many features:
  • ​​GATs, specialization, implied bounds (link to RFCs)
  • ​​Need to have something we can test to help direct further work
  • ​​Initial goal is to check some very simple Rust programs using Chalk
​​
  • ​​add non-normalized projection variant (53683)
  • ​​Estimate: 24 person-hours
  • ​​finish lowering rules (49177)
  • ​​Estimate: 16 person-hours
  • ​​implement very naive procedure to find the lowered rules for a given domain goal
  • ​​Estimate: 24 person-hours
  • ​​implement non-naive procedure for finding lowered rules for a given domain goal, integrating with incremental compilation
  • ​​Estimate: 40 person-hours
  • ​​implement unification procedures in integration
  • ​​Estimate: 24 person-hours
  • ​​implement other bits of integration (fill in panics)
  • ​​Estimate: 24 person-hours

​​preliminary support for GATs (44265)

  • ​​a lot of people have been asking for this!
  • ​​see tracking issue, …
  • ​​parsing for GATs: fix tests (comment)
  • ​​Estimated time: 24 person-hours
  • ​​improve test coverage (comment)
  • ​​Estimated time: 16 person-hours
  • ​​Lower information to HIR and propagate to the chalk rules
  • ​​Estimated time: 32 person-hours
​​